24 May 2013 14:08:10
You have to start with Moneyball. The Book of Genesis for analytics in sport; the tome which lifted the darkness on the face of the deep. Michael Lewis's book, published in 2003, was a sports-business biography that became a movement. It is now shorthand for the use of detailed statistics in sport, particularly around recruitment – these days, a great many elite sport teams accept that fracking raw data produces insights the human eye, no matter how experienced, will miss. Each year the MIT Sloan Sports Analytics conference, founded in 2006 and affectionately dubbed Geekapalooza, gets bigger and more prestigious. In America, at least, almost everyone is a Moneyballist now. You can see why. The story of how Billy Beane, and the maths wizards he appointed, transformed decades of major league baseball methodology is gripping. A tiny budget – the third smallest in the league – forced the Oakland Athletics to become revolutionaries. Instead of listening to grizzled scouts with prejudices as ingrained as the varnish on the bleachers, they trusted in statistical models and metrics. Other teams' hand-me-downs, players ignored or undervalued, were recruited. Many – surprisingly – flourished. So did the team. The As went on a 20-game winning streak, a league record.
With football, the situation regarding analytics is far fuzzier – the game is more fluid and much more complicated to track and analyse. What's more, the experts are yet to convince the masses or all of the managers. The mere mention of the word "data" makes some football fans' eyes glaze over. So in what way is "everything we know about football" wrong?
The Numbers Game begins with a question, asked by Sally, a behavioural economist and football novice, of Anderson, who played in the German fourth division before becoming a politics professor: why, if Stoke City have had repeated success with long throws, haven't other teams copied them? Anderson's answers prompt further questions and encourage the pair to chronicle the history of football analytics from pioneers such as Charles Reep – who in the 1950s published his theories in the magazine Match Analysis – to the present day, debunking several myths en route.
Pundits, armchair and professional, will find that several of their long-cherished truisms are not true at all. That a team is most vulnerable immediately after they score has been a staple of TV commentary for decades, but the data shows it to be false. These minutes are, in fact, when a team is least likely to concede. Corner kicks have little effect when it comes to increasing a team's chances of scoring, and even having more shots on goal doesn't guarantee success. Anderson and Sally examined 8,232 matches in the major European leagues from 2005-2011 and found that the team that has more shots on target only wins between 50-58% of the time, depending on the country.
The authors argue that weaker players have a bigger impact on a team's points tally than stronger ones, and claim convincingly that success in a single match is as much about luck as skill. Also intriguingly, they hold that there is a relationship between a player's disciplinary record and civil conflict in his home country, writing: "As the number of years a country has experienced of civil war goes up, so does the average number of yellow cards per player from the country."
Occasionally, however, the numbers don't tell the full story. The authors think Chelsea should have signed Darren Bent from Sunderland for £25m rather than Fernando Torres for £50m in 2011 because Bent's goals directly led to more points than those of other players in the Premier League from 2009-2011. But, as most experienced football-watchers could tell you, Bent is weak at holding up the ball and releasing others – a role he would have been required to play at Chelsea.
Anderson and Sally are adroit at guiding their readers but never patronising or boring them – which isn't easy when you're explaining the importance to football of Poisson distribution, at the centre of which is the idea that the probability of a goal being scored by a given team is the same every minute, no matter the current score or how many minutes are left. But while their writing is breezy, it sometimes overreaches: "a storm is gathering in football. It is one that will wash away all certainties and change the game we know and love" – and so on.
There is no footballing Billy Beane, which might be because top clubs are reluctant even to hint at their data discoveries in case the competition finds them out. But in reality many managers remain data-sceptics. As the authors point out, "in football nobody is quite sure what to do" with statistics. In a recent presentation to analysts at a summit in London, Anderson himself said: "I don't think football analytics has progressed as far as it could. In fact, in many ways it has stalled … It is in danger of becoming another fad." His tone resembled that of a frustrated don trying to get a talented but wayward first-year undergraduate to focus. "You are accumulating mountains of data just to sit on ... But does it change people's minds?" he asked them. "Does it help your team win? If not, you may want to do something different. Go travelling. Take up insurance. You might not want to do analytics."
The difficulty they face is indicated by a story from the book. An unnamed club asked the authors to put together a research project looking at how best to strengthen their squad and, while the board liked it, the manager was dismissive: "Stats can't tell me who to sign," he said. "They can't measure the size of a player's heart." Moneyball has a way to go in football. There are evangelicals: the analysts at the richest clubs and the academics who use tools such as "minute-by-minute bivariate probit analysis" that send you dizzy just thinking about them. But as yet few in the congregation have been listening.
